diff --git a/api/core/model_runtime/model_providers/google/llm/_position.yaml b/api/core/model_runtime/model_providers/google/llm/_position.yaml
index ab3081db38fc37..0b5e1025bfeb83 100644
--- a/api/core/model_runtime/model_providers/google/llm/_position.yaml
+++ b/api/core/model_runtime/model_providers/google/llm/_position.yaml
@@ -1,3 +1,4 @@
+- gemini-2.0-flash-exp
 - gemini-1.5-pro
 - gemini-1.5-pro-latest
 - gemini-1.5-pro-001
@@ -11,6 +12,8 @@
 - gemini-1.5-flash-exp-0827
 - gemini-1.5-flash-8b-exp-0827
 - gemini-1.5-flash-8b-exp-0924
+- gemini-exp-1206
+- gemini-exp-1121
 - gemini-exp-1114
 - gemini-pro
 - gemini-pro-vision
diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-2.0-flash-exp.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-2.0-flash-exp.yaml
index bcd59623a78e43..966617e9020d57 100644
--- a/api/core/model_runtime/model_providers/google/llm/gemini-2.0-flash-exp.yaml
+++ b/api/core/model_runtime/model_providers/google/llm/gemini-2.0-flash-exp.yaml
@@ -8,6 +8,8 @@ features:
   - tool-call
   - stream-tool-call
   - document
+  - video
+  - audio
 model_properties:
   mode: chat
   context_size: 1048576
diff --git a/api/core/model_runtime/model_providers/google/llm/gemini-exp-1206.yaml b/api/core/model_runtime/model_providers/google/llm/gemini-exp-1206.yaml
index 1743d8b9687421..7a7c361c43e18f 100644
--- a/api/core/model_runtime/model_providers/google/llm/gemini-exp-1206.yaml
+++ b/api/core/model_runtime/model_providers/google/llm/gemini-exp-1206.yaml
@@ -7,6 +7,9 @@ features:
   - vision
   - tool-call
   - stream-tool-call
+  - document
+  - video
+  - audio
 model_properties:
   mode: chat
   context_size: 2097152
diff --git a/api/core/model_runtime/model_providers/google/llm/learnlm-1.5-pro-experimental.yaml b/api/core/model_runtime/model_providers/google/llm/learnlm-1.5-pro-experimental.yaml
index 0b29814289d8b5..f6d90d52ec8820 100644
--- a/api/core/model_runtime/model_providers/google/llm/learnlm-1.5-pro-experimental.yaml
+++ b/api/core/model_runtime/model_providers/google/llm/learnlm-1.5-pro-experimental.yaml
@@ -7,6 +7,9 @@ features:
   - vision
   - tool-call
   - stream-tool-call
+  - document
+  - video
+  - audio
 model_properties:
   mode: chat
   context_size: 32767
diff --git a/api/core/model_runtime/model_providers/google/llm/llm.py b/api/core/model_runtime/model_providers/google/llm/llm.py
index 9a1b13f96f49c0..c6a452117b4a03 100644
--- a/api/core/model_runtime/model_providers/google/llm/llm.py
+++ b/api/core/model_runtime/model_providers/google/llm/llm.py
@@ -16,7 +16,6 @@
 from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta
 from core.model_runtime.entities.message_entities import (
     AssistantPromptMessage,
-    ImagePromptMessageContent,
     PromptMessage,
     PromptMessageContent,
     PromptMessageContentType,
@@ -24,7 +23,6 @@
     SystemPromptMessage,
     ToolPromptMessage,
     UserPromptMessage,
-    VideoPromptMessageContent,
 )
 from core.model_runtime.errors.invoke import (
     InvokeAuthorizationError,
@@ -342,35 +340,31 @@ def _upload_file_content_to_google(self, message_content: PromptMessageContent)
             except:
                 pass
         with tempfile.NamedTemporaryFile(delete=False) as temp_file:
-            if message_content.data.startswith("data:"):
-                metadata, base64_data = message_content.data.split(",", 1)
-                file_content = base64.b64decode(base64_data)
-                mime_type = metadata.split(";", 1)[0].split(":")[1]
+            if message_content.base64_data:
+                file_content = base64.b64decode(message_content.base64_data)
                 temp_file.write(file_content)
             else:
-                # only ImagePromptMessageContent and VideoPromptMessageContent has url
                 try:
-                    response = requests.get(message_content.data)
+                    response = requests.get(message_content.url)
                     response.raise_for_status()
-                    if message_content.type is ImagePromptMessageContent:
-                        prefix = "image/"
-                    elif message_content.type is VideoPromptMessageContent:
-                        prefix = "video/"
-                    mime_type = prefix + message_content.format
                     temp_file.write(response.content)
                 except Exception as ex:
-                    raise ValueError(f"Failed to fetch data from url {message_content.data}, {ex}")
+                    raise ValueError(f"Failed to fetch data from url {message_content.url}, {ex}")
             temp_file.flush()
+
+        file = genai.upload_file(path=temp_file.name, mime_type=message_content.mime_type)
+        while file.state.name == "PROCESSING":
+            time.sleep(5)
+            file = genai.get_file(file.name)
+        # google will delete your upload files in 2 days.
+        redis_client.setex(key, 47 * 60 * 60, file.name)
+
         try:
-            file = genai.upload_file(path=temp_file.name, mime_type=mime_type)
-            while file.state.name == "PROCESSING":
-                time.sleep(5)
-                file = genai.get_file(file.name)
-            # google will delete your upload files in 2 days.
-            redis_client.setex(key, 47 * 60 * 60, file.name)
-            return file
-        finally:
             os.unlink(temp_file.name)
+        except PermissionError:
+            # windows may raise permission error
+            pass
+        return file
 
     def _format_message_to_glm_content(self, message: PromptMessage) -> ContentType:
         """